The Plan, One Year On

On 9 April 2025, Executive Vice-President Henna Virkkunen stood before the press in Brussels and laid out a five-pillar strategy to turn Europe into the "AI Continent." The pillars covered computing infrastructure, data access, algorithm development, sectoral adoption, talent and skills, and regulatory simplification. The headline number was staggering: InvestAI would mobilize €200 billion in public and private capital for AI across the European Union, including up to €20 billion for AI Gigafactories alone.

The ambition was real. Thirteen AI Factories were already being deployed around EuroHPC supercomputers. A call for Gigafactory consortia - facilities with roughly 100,000 state-of-the-art AI chips, four times the capacity of existing AI Factories - went out the same day. A Cloud and AI Development Act was proposed to triple the EU's data center capacity over 5 to 7 years. A Data Union Strategy would create a single internal market for AI-ready data.

On paper, this was Europe doing what Europe has struggled to do for a decade: matching American-scale capital with European-scale coordination.

The Real Opportunity Is Not Where You Think

The most useful thing about the AI Continent Action Plan is not the infrastructure it builds. It is the market that reveals. The Commission acknowledged that only 13.5% of European companies were using AI. For SMEs - the backbone of the European economy, at 12.6% adoption - the number was even lower.

Read that as a founder or enterprise leader, and you should see one thing: white space. Massive, uncontested white space. Nearly 87% of European businesses have not yet adopted AI. That is not a crisis. That is the largest addressable market for AI products and services anywhere in the developed world.

A CSIS analysis published weeks after the launch made a useful distinction: Europe's path to AI competitiveness may not run through competing with American hyperscalers on raw compute. It may run through smaller, efficient models and practical application tools. When DeepSeek emerged in January 2025, European startups were already implementing similar cost-efficient training approaches. The opportunity for European builders is not to out-spend Silicon Valley. It is to out-apply it - to get AI into the hands of the businesses that need it most.

The Regulatory Landscape: What Business Leaders Need to Know Now

The AI Act entered into force in August 2024, and the regulatory picture has evolved rapidly since. Understanding where it stands today is essential for any European business building or deploying AI products.

The first obligations - prohibited AI practices and AI literacy requirements - took effect in February 2025. These are already live. If your company uses AI, your team should be trained in it. If you are using social scoring, manipulative AI, or real-time biometric surveillance without a lawful basis, you are already non-compliant.

The high-risk AI system requirements, originally due in August 2026, have been pushed back. The Digital Omnibus on AI, published in November 2025, delays these deadlines by up to 16 months and conditions enforcement on the availability of harmonized standards that are still under development. The hard backstop is December 2027 for most high-risk systems and August 2028 for AI embedded in regulated products.

The stated goal is to reduce administrative burden by at least 25% for all businesses and 35% for SMEs by 2029. For founders and enterprise leaders, this means the compliance window is wider than originally planned - but it is not indefinite.

In March 2026, the EU Council adopted its negotiating position. The European Parliament voted it through by 569 votes, with both institutions agreeing on fixed delay deadlines and adding a new ban on AI-generated non-consensual intimate content. Trilogue negotiations are now underway.

The practical reading for business leaders: Brussels heard the industry feedback and is adjusting. The Draghi competitiveness report, the Budapest declaration calling for a "simplification revolution," and direct stakeholder input all pushed the Commission toward pragmatism. The AI Act is not going away - its core protections and trust framework remain intact - but its implementation is being calibrated to match the readiness of both standards bodies and businesses.

The 86.5% - Europe's Biggest Business Opportunity

Here is the number that should frame every business conversation about Europe's AI future: 86.5% of European companies have not yet adopted AI. Among SMEs, it is 87.4%. These are not companies that rejected AI. They are companies waiting for solutions that meet them where they are - without requiring AI engineers, data scientists, or six-figure implementation budgets.

The Apply AI Strategy, launched in October 2025, is Brussels' direct response to this reality. It targets sectoral adoption across healthcare, manufacturing, public administration, robotics, and climate. It leans on the European Digital Innovation Hubs (which will refocus as AI experience centers by December 2025) and connects SMEs to the AI Factory ecosystem for compute access.

The World Economic Forum reported that fewer than 1% of companies worldwide have fully operationalized responsible AI governance, with more than 60% of European firms at the earliest maturity stages. This tells us something important: the race is not over. Most businesses everywhere are still figuring this out. European companies that move now - even modestly, even imperfectly - will be ahead of the global curve, not behind it.

Denmark ranks among Europe's top AI adopters. From Copenhagen, I can tell you that even in a mature innovation ecosystem, what SMEs and mid-market companies need most is not access to a 100,000-chip Gigafactory. They need practical tools that translate business logic into working AI-powered products. They need platforms that make deploying AI as straightforward as launching a website. The infrastructure the Commission is building matters - but it only creates value when the application layer connects it to real businesses with real problems to solve.

Three Things the Plan Gets Right - and One Critical Gap

The AI Continent Action Plan gets three things genuinely right, and European businesses should take advantage of all three.

First, the Data Union Strategy. Europe's fragmented data landscape has been a real barrier. Creating Common European Data Spaces and establishing cross-sector data-sharing norms will take years, but when it arrives, European companies that have built data-aware products will have a structural advantage over competitors who treated data as an afterthought.

Second, the talent pipeline. The AI Skills Academy, international recruitment pathways, and Marie Sklodowska-Curie fellowships address a genuine constraint. Europe trains excellent AI talent. Keeping that talent requires competitive conditions - and the Commission is putting mechanisms in place. For enterprise leaders: this means a growing pool of AI-literate professionals entering the European market over the next three to five years. Plan your hiring accordingly.

Third, the regulatory sandbox model. Both the national sandboxes (operational by August 2026) and the new EU-level sandbox for general-purpose AI models provide companies with a structured way to test innovations under regulatory guidance. If you are developing AI products that might touch high-risk categories, apply for sandbox access. It is one of the most practical resources the Commission has created.

The gap worth noting: the plan builds infrastructure at the top (compute, data centers, Gigafactories) and provides regulatory relief at the bottom (simplification, extended deadlines, SME exemptions). But the middle layer - the application tools, platforms, and products that connect infrastructure to actual business adoption - receives comparatively little structural support. The Apply AI Strategy gestures toward it through sectoral dialogues and Digital Innovation Hubs. But the real work of closing the adoption gap will be done by builders: by the companies and platforms that make AI useful for the 86.5% who have not started yet.

What European Business Leaders Should Do Now

The regulatory window is open. Compliance deadlines have been pushed back. Sandbox access is expanding. SME exemptions are broadening. The infrastructure pipeline is being built. For the first time in years, the policy, capital, and technology environments are aligned in Europe's favor.

Here is what matters for the next twelve months. The Digital Omnibus trilogue negotiations will finalize compliance timelines - follow them closely if you are building high-risk AI systems. The drafts of the Multiannual Financial Framework 2028-2034 will reveal how much sustained public investment follows the initial commitments. The Gigafactory consortia will begin to take shape, and the Digital Fitness Check will signal how far Brussels is willing to go on ongoing simplification.

But do not wait for any of that to start. The companies that will define Europe's AI future are not waiting for perfect regulatory clarity or fully operational Gigafactories. They are building now - deploying AI into European verticals, solving European business problems, and serving the enormous market of enterprises that have not yet taken their first step.

The AI Continent Action Plan gave Europe the right ambition. It committed real capital. It opened the regulatory door. Now it is on us - founders, enterprise leaders, product builders - to walk through it.